Browse Prior Art Database

Key Structure Management of Data Within a Set of Members in a Partitioned Data Set

IP.com Disclosure Number: IPCOM000042850D
Original Publication Date: 1984-Jun-01
Included in the Prior Art Database: 2005-Feb-04
Document File: 3 page(s) / 93K

Publishing Venue

IBM

Related People

Florio, A: AUTHOR [+2]

Abstract

This article presents a method of storing named data items within members in a partitioned data set. It provides an ability to significantly increase the number of unique named data items in that data set using a minimum number of members. At the same time, this method minimizes the storage and input/output overhead required to locate each unique data item and minimizes the number of members accessed at any one time to facilitate recovery of the data items in case of failure. A key structure management algorithm is used to store many different named data items in each of a set of members. Each key includes a member number, a record chain number, and a record number. Name lists and lists of available keys are stored in the first part of the first member.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 3

Key Structure Management of Data Within a Set of Members in a Partitioned Data Set

This article presents a method of storing named data items within members in a partitioned data set. It provides an ability to significantly increase the number of unique named data items in that data set using a minimum number of members. At the same time, this method minimizes the storage and input/output overhead required to locate each unique data item and minimizes the number of members accessed at any one time to facilitate recovery of the data items in case of failure. A key structure management algorithm is used to store many different named data items in each of a set of members. Each key includes a member number, a record chain number, and a record number. Name lists and lists of available keys are stored in the first part of the first member. When accessing a data item by name, the name list is searched for the given name to get the associated key. The key then points to the member in the partitioned data set, the record chain within that member, and the record within that record chain for the associated data item. As shown in the figure, member 1 contains thirty-two different name lists which are retrievable by a unique pointer. The unique pointer consists of two characters: a fixed first character (F1) indicating that it is a name list, and a second character (consisting of a number from 1 to 32) identifying the name list, itself. Each of the 32 name lists contain a set of name entries. Each name entry has the alphanumeric name and an associated key to the named data item. The key consists of three characters. The first character is the member number, the second character is the record chain number, and the third character is the number identifying the records in the chain that make up the data item. In the example in the figure, the key for data item "NAME4" is 02 01 01. 02 points to member 2, 01 points to the first record chain in member 2, and 01 indicates that "NAME4" is the first record in record chain one in member 2. Member 1 also contains free element lists describing the unique keys available to be assigned to a data item within each member. Accordingly, there are n lists matching the number of members involved within the partitioned data set. The free element lists are retrievable by a unique pointer consisting of two characters: a fixed first character of (F2) indicating that it is a free el...